ABSTRACT
Real-time stereo matching with high accuracy is a dynamic research topic; it is attractive in diverse computer vision applications. This paper presents a stereo-matching algorithm that produces high-quality disparity map while maintaining real-time performance. The proposed stereo-matching method is based on three per-pixel difference measurements with adjustment elements. The absolute differences and the gradient matching are combined with a colour-weighted extension of complete rank transform to reduce the effect of radiometric distortion. The disparity calculation is realized using improved dynamic programming that optimizes along and across all scanlines. It solves the inter-scanline inconsistency problem and increases the matching accuracy. The proposed algorithm is implemented on parallel high-performance graphic hardware using the Compute Unified Device Architecture to reach over 240 million disparity evaluations per second. The processing speed of our algorithm reaches 98 frames per second on 240 × 320-pixel images and 32 disparity levels. Our method ranks fourth in terms of accuracy and runtime for quarter-resolution images in the Middlebury stereo benchmark.
Acknowledgement
The authors extend their appreciation to the Deanship of Scientific Research at King Khalid University for funding this work through Small Groups. (Project under grant number RGP.1/367/43).
Disclosure statement
No potential conflict of interest was reported by the author(s).
Additional information
Funding
Notes on contributors
Mohamed Hallek
Mohamed Hallek received his Ph.D Degree in Micro-electronics from the Faculty of Sciences of Monastir, Tunisia, in 2022. He is a member of the Laboratory of Electronics and Microelectronics and a Teacher researcher at the Faculty of Sciences Monastir, Tunisia. His fields of interest are pattern recognition, image and video processing in graphic processors and stereo-matching algorithms.
Randa Khemiri
Randa Khemiri received her Ph.D Degree in Micro-electronics from the Faculty of Sciences of Monastir, Tunisia, in 2017. She is currently an assistant professor at the Higher School of Sciences and Technology of Hamman Sousse (Tunisia). Her research includes image and video processing in graphics processors, motion tracking and pattern recognition, medical imaging, deep learning, and circuit and system design.
Ali Aseere
Ali Aseere received his Ph.D degree in Computer Science from the University of Southampton, United Kingdom in 2012. He has published many research papers in refereed International Journals/Proceedings/Books. He is currently working as an associate professor at the Department of Computer Science, King Khalid University and he also holds the Dean position at the College of Computer Science. He served as session chair and organizing committee member of various conferences. His research areas are Intelligent and Multi-Agent Systems, Agent-based models and Agent mining.
Abdellatif Mtibaa
Abdellatif Mtibaa received his Ph.D. degree in Electrical Engineering at the National School of Engineering of Tunis. Since 1990, he has been Assistant Professor in Micro Electronics and Hardware Design in the Electrical Department at the National School of Engineering of Monastir. Since 2007, he has been a professor at the Electrical Engineering Department at the ENIM. His research interests include high-level synthesis, rapid prototyping and reconfigurable architectures for real-time multimedia applications.
Mohamed Atri
Mohamed Atri is Professor in the Department of Computer Engineering at the College of Computer Science, King Khalid University, Saudi Arabia. He received his Ph.D. Degree in Microelectronics from the University of Monastir, Tunisia in 2001 and the HDR degree in 2011. He is currently a member of the Laboratory of Electronics & Microelectronics, Faculty of Sciences of Monastir. His research interests include AI, Circuit and System Design, Image processing, and Network Communication. He is the co-author of more than 200 publications including refereed journals, conference papers and book chapters.